The increasing needs of telecommunications capacity requires the use of higher network bandwidths. Increasing the bandwidth in optical telecommunications systems is mainly performed by increasing the number of channels, and better utilization of the large bandwidth of optical fibers. Increasing the number of channels per fiber is achieved by assigning each channel a separate wavelength. This is known as Wavelength Division Multiplexing, or WDM. Such systems are typically composed of a large number of wavelength dependent components, starting with the laser transmitter source, continuing through optical amplifiers, wavelength selective elements or wavelength discriminating devices, such as filters, multiplexers and demultiplexers, and ending at the receivers. In order to properly utilize the large resources of such a system, proper optimization is necessary to avoid effects between channels such as crosstalk, and to optimize performance in terms of power budget. Because of the dense spectral utilization of such systems, wavelength optimization is one of the most important considerations of such a system. Mismatch between the intended wavelengths of transmitters in a system, for instance, will cause mutual interference with other transmitters, known as crosstalk, which will express itself in data loss and severe power penalties in the relevant channels in the system. Therefore, it is very important in such systems to ensure that transmission wavelengths are optimized with respect to the system requirements, and that methods are available to monitor this continuously during the lifetime of the system, without significantly interfering with system operation.
Furthermore, the total spectral response of the system depends on many wavelength sensitive components, and therefore the total system response is prone to variation, due both to initial component tolerance, and to component aging thereafter. Even though great effort may be made to standardize all systems to specific wavelengths, such as the ITU-T wavelength recommendations, since each optical system may be slightly different in its spectral response, this cannot always be simply achieved.
Tunable semiconductor laser diodes (TSLD's) are key components in the implementation of such optical communication systems, since it is necessary to provide the ability to dynamically allocate wavelengths in a WDM system by tuning the transmitter wavelength to the wavelengths of the target receivers. Lasers can emit at wavelengths that have a round trip phase change through the cavity from end to end of 2 kπ (k in an integer). These wavelengths define the Fabry-Perot modes of the cavity. In order to ensure single longitudinal mode operation of the laser, the net gain of the preferred lasing mode must be higher then that of the other Fabry-Perot modes. In a TSLD, cavity mode discrimination is achieved using wavelength selective structures such as reflectors, sampled reflectors (SG), super structure reflectors (SSG) and couplers.
The currently most widely used TSLD's use, as the tuning mechanism, the free plasma effect in these structures. By injecting current into these structures, a change in the optical refraction index is invoked, thus changing their optical properties and ultimately changing the wavelength of the laser. Such TSLD's provide fast tuning capabilities together with accuracy, stability and repeatability. The number of structures or sections in such a laser determines the number of input currents that have to be provided in order to tune the laser.
When the laser is correctly tuned, all of these sections provide the maximal transmission at the desired wavelength and attenuate other modes, thus ensuring single mode operation. If the laser is not correctly tuned, i.e. if the wavelength selective sections are not all aligned to a common wavelength, or the phase current is not adjusted to meet the required phase condition at the desired wavelength, the laser may operate in an unstable mode, or not operate at all. An arbitrary set of tuning currents usually results in no lasing power. This is why a multi-dimensional scan of the tuning currents, in an attempt to find a favorable combination, generally yields poor results.
Characterization of a tunable, multiple section semiconductor laser is generally understood in the art to mean a process in which the tuning currents necessary to tune the laser to given wavelengths are determined. The characterization procedure presents a problem in laser manufacture. This information cannot be determined during production, and is different for each individual laser due to microscopic differences between the lasers.
Current techniques for the characterization of multiple section lasers, such as DBR, SG-DBR, and GCSR lasers, use methods largely based on trial and error. A set of currents is introduced to the laser, and the wavelength and optical power are measured. The set is slightly changed and the process is repeated. As there are generally three or four different tuning currents, depending on the number of sections, this prior art process involves a three-dimensional or four-dimensional scan of the input currents to the laser. In this way the tuning currents for different wavelengths are obtained. This method is very time and labor consuming, currently taking several hours per laser, and adds significant additional costs to the laser. More recent developments of this basic method use what could be termed as intelligently directed trial and error methods, but the characterization procedure is still a time consuming and inefficient process, since it provides a large quantity of unneeded data for areas of operation that are not necessarily used in the application envisaged. Such characterization techniques generally focus on full characterization of the laser regardless of the area of its intended use.
A recent method of characterization is described in the PCT Patent Application by B. Broberg et al, for “Method of optimizing the operation points of lasers and means for carrying out the method”, published as No. WO 99/40654 and hereby incorporated by reference in its entirety. A method is described therein whereby a sensing device attached to the laser output senses power discontinuities occurring at mode jumps of the laser, and calculates a series of mode planes of allowed operation away from these mode jumps. These mode planes are stored in a control system, which then ensures that the currents to each of the laser sections maintain operation of the laser on these mode planes. The outcome of the characterization described therein would appear, to the best understanding of the current applicants, to be another method for ensuring operation of the laser away from the well-known discontinuities occurring at mode jumps. Furthermore, although this method does reduce the magnitude of the characterization problem in comparison with previous methods, which characterize the laser even in forbidden areas of operation, the method still requires a large number of wavelength measurements to be performed within each mode plane, in order to relate operational points on the mode planes to actual wavelengths. These measurements methods, typically performed using an optical speck analyzer, are labor intensive and time-consuming. Furthermore, this approach does not seem to relate the wavelength characterization directly to the application in hand, but rather provides a cross reference between tuning currents and values of wavelength, which are then referred to those values of wavelength required by, or known from the application.
Another recent method of laser characterization is given in the article entitled “Novel mode stabilization scheme for widely tunable lasers” by G. Sarlet et al, published in the Proceedings of the European Conference on Optical Communications (ECOC) 1999, pp. 128-129, hereby incorporated by reference in its entirety. In this article, the authors describe a similar method of characterizing an SSG-DBR laser, in which a several scans are performed of the front SSG, rear SSG and phase section. In effect, three separate sets of two dimensional scans of two of the three sections are performed sequentially, thereby covering the entire operating envelope. This method too has the disadvantage of requiring many measurements, and is very time and labor intensive. Furthermore, it leads to the establishment of stable working points, but having no relation to the actual wavelength at the laser output. In order to relate to actual emitted wavelengths, an additional large number of measurements have to be performed, as in the previously mentioned prior art, using, for instance, an optical spectrum analyzer.
There therefore exists an urgent need for a method of precise characterization of TSLD's, which can be performed simply, speedily and at low cost, and preferably without special test equipment, and in particular, which optimizes the laser wavelengths to the requirements of the system in which it is used. Such a characterization would also assist in optimizing total performance of the system in terms of wavelength dependency of its components. The availability of such a method would enable more cost-effective, widespread use of these devices, and better utilization of available system wavelength resources.
The disclosures of each of the publications mentioned in this section and in other sections of the specification, are hereby incorporated by reference, each in its entirety.